The metacommunity is an emergent concept that considers the impact of the exchange of species in heterogeneous environments (Leibold et al., 2004; Urban & Skelly, 2006). Temporary ponds, which are characterised by annual inundation-desiccation cycles (Williams, 1997), are ideal model systems to study metacommunity ecology given their simple structure, local abundance, and occurrence in pond networks that demonstrate clear environmental gradients (Vanschoenwinkel et al., 2007; Pandit et al., 2009). Although temporary ponds are widely distributed worldwide (Williams et al., 2001), their high biodiversity contrasts with their sensitivity and vulnerability to external perturbation, which has led to great interest in their conservation over the last few years (Williams et al., 2001; Zacharias et al., 2007; Céréghino et al., 2008). In addition, temporary ponds harbour singular flora and fauna that are often exclusive or infrequently found in permanent ponds (Collinson et al., 1995; Williams, 1997; Céréghino et al., 2008). In particular, their singular macroinvertebrate species can adjust their life cycles to the annual period of pond inundation (hydroperiod), re-starting community assembly after each year's initial inundation (Bazzanti et al., 1996; Boix et al., 2004; Florencio et al., 2009).
In metacommunity ecology, β-diversity, which is the variation in species composition among sites in a geographical area (Legendre et al., 2005; but see e.g. Tuomisto, 2010; Anderson et al., 2011), is a key concept for understanding ecosystem functionality from a management and conservation perspective. In pond networks, environmental heterogeneity has been revealed as crucial in supporting high biodiversity (Urban, 2004; Jeffries, 2005) and also in driving patterns of nested biodiversity, in which species-poor sites contain subsets of species-rich sites, particularly in those systems with good conservation status (Hylander et al., 2005; Florencio et al., 2011). Hence, the study of those species that depart from the expectations of nested biodiversity patterns, which occur more or less frequently than would be predicted in a nested system (termed idiosyncratic), is currently receiving great interest in applied ecology (e.g. Florencio et al., 2011). To better understand the ecological processes maintaining high ecosystem diversity, β-diversity should be partitioned between (i) the β-diversity associated with non-random species loss in nested systems; and (ii) the β-diversity associated with true species replacement (Baselga, 2010). It is essential to disentangle the problem whether β-diversity is driven by species replacement or nestedness to make appropriate conservation decisions. If the former is the driver, it would prioritise the conservation of a large number of sites with variable richness and environmental conditions, while the latter would prioritise the conservation of the richest sites (Baselga, 2010).
One of the main debates in metacommunity ecology involves the relative importance of deterministic, niche-based process (e.g. environmental filters) versus stochastic ecological process (e.g. dispersal filters) in community assembly (Chase & Myers, 2011). Water chemistry and the physical characteristics of ponds each have an important influence on macroinvertebrate composition and abundance in wetlands (Wissinger, 1999; Williams, 2006). Conductivity is one of the most frequent chemical descriptors of macroinvertebrate communities (Garrido & Munilla, 2008; Waterkeyn et al., 2008). In particular, acidic water has negative effects on macroinvertebrate species diversity (Radke et al., 2003). Although nutrient concentrations have controversial effects, they usually negatively impact species occurrences at high levels (Declerck et al., 2005). Applying the theory of island biogeography (MacArthur & Wilson, 1967) to lakes and ponds, high macroinvertebrate and plant species richness is harboured in large ponds (Friday, 1987; Nicolet et al., 2004). Interpond distances can also affect the incidence of species in particular pond assemblages as a result of species dispersal limitations (Briers & Biggs, 2005; Sanderson et al., 2005).
We explored the main drivers of β-diversity and community structure in a macroinvertebrate metacommunity in a pond network of excellent conservation status. This is a highly dynamic system in which thousands of ponds fill and desiccate annually, with only a few ponds retaining water during the summer. The novelty of our study resides in the fact that we obtained comparable data on macroinvertebrates in 80 ponds distributed across an extensive area. We hypothesised that (i) there is high biodiversity in the macroinvertebrate metacommunity, with species replacement and nestedness being the main contributors to β-diversity; (ii) environmental variability is key in maintaining such high macroinvertebrate diversity in the pond network; and (iii) both random (i.e. dispersal) and deterministic processes (i.e. environment) are operating together in the macroinvertebrate assembly. To evaluate these hypotheses, we used data from 80 ponds, collected over a single season, to analyse (i) if β-diversity was mainly sustained by nestedness or by species replacement, and (ii) if spatial connectivity and environmental variability had an important influence on macroinvertebrate structure and β-diversity.